503
THEHYDROGENATION of ETHANE ON COBALT CATALYSTS
Feb., 1939
TABLE IV ACTIVITYCOEFFICIENTS m
0 . 05
,079 .10 .14 .20
.30
.40 .50 .60 .70 .80 1.00
“/‘no
0.811 .782 .766(0) .744(0) .721(0) .697(0) .682(0) .672(0) .665(0) .661( $1) .657(0) .655(+1)
-/‘no
7’80
Y’70
0.807 ,777 .762($1) .739( 4-1) .717(+1) .691(0) .676(+1) .667 ( 4-2) .659( 4-2) .654( 1) .651(+1) .648(+1)
+
0.803 ,773 .757(0) .734(0) .711(0) .686(0) ,671(4-1) .660(0) .653( 1) .648(+1) .644( $1) .641( 1)
+ +
Y‘im
0,799 ,768 .752(0) .729( fl) .705(0) .679( - 1) .663( - 1) .653(-1) .645(0) .640(0) .636(0) .632(-1)
0.794 ,703 ,746 (0) .722(0) .698(0) .672(-1) .655( - 1) .644( -2) .636(-1) .631( - 1) .627( - 1) .622( -2)
the activity coefficients of sodium chloride de- of 8 a t 60’ were slightly too high. Thisis quite termined from the boiling point elevation and reasonable in view of the greater uncertainty of from electromotive force measurements, are the measurements a t the lower pressures.13 plotted against the temperature. To save space Summary each curve has a different reference point, the 1. The boiling point elevations of aqueous ordinate figures give the values of y at 0’ for the corresponding curve. The diameter of each sodium chloride solutions have been measured circle represents a variation in y of 10.001. a t reference temperatures of 60, 70, 80, 90, and An inspection of Fig. 2 shows an excellent agree- looo, and a t a number of concentrations between ment between the results obtained by the two 0.05 and 1.0 M . 2 . The activity coefficients of the solute have methods. While the values of y a t 60’ appear to be too high by 0.001 to 0.002 a t concentra- been calculated. 3. A comparison of the results obtained from tions above 0.2 M, these deviations are not greater than the expected experimental error. the boiling point rise with those obtained from However, the fact that they are all in the same e. m. f. measurements is given. direction would indicate that the observed values NEWHAVEN, CONN. RECEIVED JULY 29, 193X
[CONTRIBUTION FROM THE
FRICK CHEMICAL LABORATORY, PRINCETON
UNIVERSITY]
The Hydrogenation of Ethane on Cobalt Catalysts BY ELLISON HALLTAYLOR AND HUGHS. TAYLOR The immediate precursor of the present study was the work of Morikawa, et al.,1)2i3 on the activation of the C-H and C-C bonds in hydrocarbons a t the surface of a nickel catalyst. By studying two reactions of the same two reactants on the same surface, these authors were able to give to the term “specificity” a concrete meaning. It became immediately a problem of interest to extend the study to other catalysts, in order to see how the factors affecting such reactions would vary from catalyst to catalyst. The hydrogenation-decomposition of ethane gave results of more (1) Morikawa Benedict and Taylor. Txrs JOURNAL, 68, 1445 (1936). (2) I b i d . , 58, 1795 (1936). (3) Morikawa, Trenner and Taylor, i b i d . , 59, 1103 (1937).
interest than the exchange, first because of the abnormally high hydrogen dependence and apparent activation energy, and second because it involved breaking a carbon-carbon bond, a step of more general importance in the transformations of other hydrocarbons than breaking a carbon-hydrogen bond. It was therefore decided to study that reaction on a different catalyst. Cobalt presented itself as an interesting catalyst for such a study. One reason for the selection of cobalt was its rising industrial importance as a catalyst in the Fischer process4 It was therefore decided to use first a catalyst typical of those actually used in that process, one of (4) Franz Fischer, Brenn4o.f Chemic, 16, 1 (1935)
504
ELLISON HALLTAYLOR AND HUGH s. TAYLOR
cobalt, thoria, copper and kieselguhr. Work in an entirely different direction led to the choice of a series of cobalt-magnesia catalysts for the other type to be studied. A series of papers by J. R. L e ~ i s , ~by . ~ JLewis and Taylo$ and by Taylor, Joris and V e r h u l ~ t had ~ , ~ revealed ~ the stability of dilute copper-magnesia catalysts and the high degree of dispersion of the copper as measured by adsorption, reactivity after high temperature treatment and by X-ray studies. Results, as yet unpublished, indicate that cobalt can be similarly very finely dispersed by coprecipitation with magnesium hydroxide.
Experimental Details Gases.-Hydrogen was prepared by electrolysis of sodium hydroxide solution. Traces of oxygen were removed by passing the gas over a red-hot spiral of platinum wire, the hydrogen being dried before use by slow passage through a trap immersed in liquid air. The ethane used was prepared by hydrogenation of ethylene on a nickel catalyst followed by combustion of excess hydrogen on copper-copper oxide and by fractional distillation t o remove methane and nitrogen. Before use, the ethane was further purified by repeated condensation in liquid air and evacuation of any residual permanent gases. Catalysts.-The cobalt, copper, thoria catalyst no. 2 was prepared from 0.315, 0.035 and 0.070 mole of the respective nitrates dissolved together in distilled water and diluted to 0.2 molar in total solutes. A 0.2 molar solution of 0.59 mole of sodium carbonate was used for precipitation. To the solution of the nitrates 186 g. of kieselguhr was added and brought into suspension by vigorous stirring. The sodium carbonate solution was then added a t the rate of 50-70 drops per minute, with continued stirring. Precipitation was incomplete at the equivalence point and so the whole of the carbonate (supposedly 0.1 mole excess) was added. The precipitate was washed nine times by decantation, filtered with suction and dried a t 110". A weighed portion of the catalyst, in the reaction vessel, was evacuated twelve hours a t 255" and reduced in hydrogen a t that temperature for 150 hours and then a t 350" for thirty hours. The cobalt-magnesia catalysts, 4A, B, C, D, E, were prepared from stock solutions of cobalt and magnesium nitrates and potassium hydroxide. Five solutions, 500cc. volume, each containing 0.03 mole of cobalt and 1, 2, 3, 4 and 5 times 0.03 mole of magnesium, respectively, were precipitated by addition of hydroxide at 3 cc. per minute with vigorous stirring continued in each case for one hour after completion of precipitation. The precipitates settled for twenty-four hours, were filtered with suction and washed four times each by stirring the filter cake into (5) Lewis, J . Phys. Chcm., 32, 243 (1928). (6) Ibid., 35, 915 (1931). (7) l b i d . , 37, 917 (1933). ( 8 ) Lewis and Taylor, THIS JOURNAL. 60, S i 7 (1938). (9) Taylor and Joris, Bull. SOC. chim. Bclg., 46, 241 (1937). (10) Joris, Verhulst and Taylor, ibid., 46, 393 (1937).
Vol. 61
a paste with a little water and then adding one liter of water to form a smooth suspension, finally filtering with suction. After the fourth washing, the supernatant liquids gave only a faint potassium flame comparable with that given by 10-6 molar potassium solutions. The dried catalysts, weighed into the reaction vessels, were reduced in streams of hydrogen, a t a rate of 750 cc. per hour, the temperature of the catalyst being raised over a period of four hours to 425' and held there for another four hours of reduction. Analyses of catalysts nos. 2, 4A and 4E showed 22.7, 38.3 and 14.0% cobalt in close agreement with the expected compositions. Catalyst no. 2 was gray-pink when dried before reduction and yielded a black reduced product. The dried catalysts nos. 4A to 4E varied progessively in color from black through brown to light tan, which color-sequence was maintained after reduction. Apparatus and Procedure.-The apparatus consisted of the following parts: (1) apparatus for preparing, purifying and storing the gases, (2) apparatus for measuring the gases, (3) a reaction vessel, with necessary auxiliary equipment, (4) an automatic pump of modified T6pler type for removing the gases rapidly from the reaction vessel and handling them during the analysis, (5) apparatus for analyzing the gases and ( 6 ) a pumping system for evacuating the catalyst and the rest of the apparatus. The reaction vessels were Pyrex tubes about 22 mm. in external diameter, 25 cm. long and 115 cc. in volume. The vessel was initially connected, top and bottom, through capillary tubing to a stopcock. After reduction, the bottom tube was sealed off close to the vessel. The catalyst was contained in a small glass cup a t the bottom of the vessel. This kept it from being blown about while hydrogen was being passed through for reduction. The catalyst was protected from mercury vapor by a U-tube filled with copper shot and cooled in a mixture of solid carbon dioxide and acetone placed immediately beyond the inlet stopcock. A manometer and gas reservoir, which consisted of a vertical capillary tube of barometric height attached at the bottom to a 125-cc. bulb served for pressure measurements. During a run, the mercury stood high in the capillary tube. The reservoir was used to hold the gas mixtures before they were put onto the catalyst after an analysis. The automatic Topler pump, which consisted of a vessel of about 600-cc. volume, was operated by electrical contacts and a vacuum tube relay to cause mercury to pump gases a t a rate of one or two strokes per minute, introducing them to or removing them from the reaction tube, the burets or the analytical system. During the combustion of hydrogen, in the analytical operation, a slower rate of pumping was used. The pump caused the gases to circulate over a copper oxide tube heated electrically to 300". The water formed was removed in two traps immersed in a solid carbon dioxide-acetone mixture. The combustion process was repeated until constant volume was obtained. The method of analysis was checked occasionally by. fractionation of the residual hydrocarbons, which was fairly satisfactory in mixtures rich in methane. A two-stage mercury vapor pump backed by a Hyvac oil pump comprised the evacuation system. Catalyst
Feb., 1939
THEHYDROGENATION OF ETHANE ON COBALTCATALYSTS
no. 2 was evacuated a t 350" and catalysts 4 A 4 E a t 425" between runs for a t least eight hours. The reaction temperatures were produced by vapor baths of aniline, 184", naphthalene, 218", diphenyl, 255", and a-bromonaphthalene, 280". The experimental procedure was normally as follows: hydrogen and ethane samples to give a desired gas mixture were mixed, introduced to the catalyst and the pressure noted. Adjusting the pressure required from one to two minutes, but the pressure was close to the desired value within about fifteen seconds after the start of the run At ten-minute intervals throughout the run, the gas was removed partially from the reaction vessel A t the end of a given interval of reaction, the gas mixture was removed rapidly by expansion into evacuated burets of 500 and 250 cc. volume, respectively, and then into the automatic Topler pump which was then operated for fifteen minutes. I n a typical run 98% of the gas was thus removed within one-half minute. In every case, from 0.1 to 1 cc. of gas could not be removed from the catalyst by this procedure (see later). The gas so collected was analyzed as already described,* and, after the analysis, a n amount of hydrogen equal to that removed by the combustion was added to the residual hydrocarbons and the reaction procedure repeated. By this method, the pressure a t the end of a run could be duplicated to well within 5%.
Calculation of Results.-From the analytical data, corrected to 0" and 760 mm., the amount of ethane decomposed during each reaction stage was calculated by subtracting the initial volume of hydrocarbons from the final volume of hydrocarbons in that stage. The difference between the initial hydrogen volume and the volume of hydrogen burned served as a check, since the hydrogen consumed should be equal to the ethane decomposed. These calculations follow, obviously, from the equation for the assumed reaction C2He Hz = 2CH4. Two assumptions are involved in the calculations, (a) that the reaction occurring is the hydrogenation of ethane, (b) that all the gas left on the catalyst after removal of the sample in the standard way was hydrogen. The following reactions might conceivably occur in addition to the reaction assumed :
505
a volume increase, which was never observed, and in a deposition of carbon. No large fall in activity was observed as is the case on nickel when carbon deposition occurs, If reaction ( 5 ) occurred a t an appreciable rate i t would yield a larger amount of methane than would be calculated from the combustion analysis. In no case was more methane found by fractionation than was calculated from the combustion analysis. The only remaining reaction is the hydrogenation of ethane to methane and so i t is justifiable to base calculations on i t as the sole appreciable reaction. This reaction requires that the ethane decomposed should equal the hydrogen used. If i t is assumed that all the gas not recovered is hydrogen, then the calculated amounts of ethane decomposed and hydrogen consumed are equal, in all cases, to well within the experimental error. If it is assumed that part of the gas not recovered is ethane, then the calculated amounts of ethane and hydrogen consumed are not in agreement, thus violating the requirement of the only possible appreciable reaction. Experimental Results
The experimental results can be presented conveniently in the form of seven diagrams, the first four dealing with the effects of temperature and gas composition on the rate of reaction on the cobalt-copper-thoria catalyst no. 2 , the last three the effect of the same variables on three cobaltmagnesia catalysts 4A, 4B and 4E. The experiments were performed in such an order as to repeat, finally, the initial experiment in the group. This procedure would have revealed any marked change in catalyst activity. The initial pressure for each experiment was chosen to give the same partial pressure of ethane, 73 mm., in all the experiments. Figure 1 shows the effect of temperature, using catalyst no. 2, on a 2H2:IC2H8mixture at 255 and 280'. Curves 1 and 3 show the reproducibility CzHe = C2Hi Hz (1) a t 255'. Figure 2 presents, for the same catalyst, 3CzHs = 2 C a ~ 3Hs (2) CzHa = 2C 4- 3Hz (3) the effect of variation in composition a t 280°, 2C2H8 = C,HI, H? (4) curves 1 and 4 for a 2 : l mixture and curves 2 and 2CzHs = CaHs + CEIi 0) 3 for a 3:l mixture. Fig. 3 presents data a t Of these five, the only ones that can take place to 255' for variation in gas composition on the same an appreciable extent are reactions (3) and ( 5 ) ) catalyst. Curves 1 and 5 are for a 2:l mixture, all others being eliminated by equilibrium con- curves 2, 3 and 4 for a 1.25:l mixture. Figure 4 siderations. Reactions (3) and ( 5 ) can be elimi- extends. this variable from a 2:l mixture (curves nated on kinetic grounds. The former results in 1 and 3 ) to a 3 : l mixture (curve 2). Figures 5 , 0
+
+
+ +
AND HUGHS. TAYLOR ELLISON BALLTAYLOR
506
Vol. 61
in the case of curve 3, Fig. 5 , where the temperature is 184' for a 2:l mixture."
0
2 4 6 Time in hours. Fig. 1.-Effect of temperature: Co-Cu-ThOz catalyst; curves 1 and 3 a t 255', curve 2 a t 280".
and 7 present in each case the data for temperature and composition variables for catalysts 4A, 4B and 4E. The mixtures are in each case 2 : l and 3:1, the temperatures are 218 and 255' save
4 6 Time in hours. Fig. 3.-Effect of reactant composition a t 255' on Co-Cu-ThOz catalyst: hydrogen-ethane ratio 2:l in curves 1 and 5, and 1.253 in curves 2 , 3 and 4. 0
In Table I is presented a summary of all the experimental results, in which table the initial rate of methane formation per hour is expressed in the fourth column and the derived rates, per
0
0
4 6 Time in hours. Fig. 2.-Effect of reactant composition at 280" on Co-Cu-Tho2 catalyst : hydrogen-ethane ratio 2:l in curves 1 and 4,3:l in curves 2 and 3. 2
2
2
4 6 Time in hours. Fig. 4.-Effect of reactant composition a t 255' on Co-Cu-Tho2 catalyst : hydrogen-ethane ratio 2:l in curves 1 and 3, and 3:l in curve 2. (11) The actual numerical d a t a from which these curves were constructed are contained in t h e P h . D . thesis of Ellison H. Taylor, Princeton University, 1938, a copy of which, in t h e University Library, may be obtained, on loan, through Inter-library Loan Service.
THEHYDROGENATION OF ETHANE ON COBALT CATALYSTS
Feb., 1939
507
gram of catalyst and per gram of cobalt in the catalyst, are given in the fifth and sixth columns. TABLEI HYDROGENATION OF ETHANE(73 MM.) Initial Temp.,
Catalyst
PH1,
mm.
OC.
2 2 2 2 2 4A 4A 4A 4B 4B 4c 4E 4E
255 255 255 280 280 184 218 218 218 218 255 218 255
92 147 220 147 220 147 147 220 147 220 147 147 147
TO METHANE Avera e initial rate-# per hour Per g. Per g. Actual cat. co
19 13 11 46 35 4 47 27 14 8 24 2 20
16 11 9 38 29 6 66 38 36 20 61 5 49
70 48 40 167 130 15 173 99 99 57 200 35 350
This table immediately reveals the superiority per gram of catalyst or of cobalt, for the conversion of ethane to methane, of cobalt dispersed in magnesia over cobalt dispersed with copper and thoria on kieselguhr.
0
2
4 6 Time in hours. Fig. 6.-Effect of temparature and reactant concentration on lCo-2MgO catalyst: curves 1 and 3 a t 218" and H2:C2H0 = 2 : l ; curve 2 a t 218" and 3:l ratio.
to the initial rates of runs made a t different temperatures but with the same partial pressures of reactants. Also, the dependence of the reaction rate on hydrogen concentration was calculated by applying the equation: rate = k &2p$2Hs&Ha to the initial rates of runs with different initial partial pressures of hydrogen, but with constant partial pressure of ethane and a t the same temperature. The reaction was assumed to be firstorder with respect to ethane (y = 1) and unaffected by methane ( z = o).
2 4 6 Time in hours. Fig. 7.-Effect of temperature and reactant concentration on 1Co-5MgO catalyst: curve 1 a t 218' and H ~ : C ~ H=B 2 : l ; curves 2 and 4 a t 255" and 2:l ratio; curve 3 a t 255' and 3:l ratio. 0
0
2
4 6 Time in hours. Fig. 5.-Effect of temperature and reactant concentration on ICo-1MgO catalyst: curves 1 and 4 a t 218" and Hz:CZHB= 2:l; curve 2 a t 218' and 3:l ratio; curve 3 a t 184" and 2:l ratio.
The apparent activation energy for each catalyst on which sufficient data were available was calculated by applying the Arrhenius equation
With the aid of the calculated activation energies the initial rates a t a single temperature, 25j0, on the different catalysts have been calcdated to facilitate comparison. These derived data are presented in Table 11.
ELLISON HALLTAYLOR AND HUGHS. TAYLOR
508
Vol. 61
TABLE I1 synthesis activity. It was hoped that the hydroINITIAL RATESAT 255 O, APPARENT ACTIVATION ENERGIES genation-decomposition of ethane might serve AND HYDROGEN DEPENDENCE OF METHANEFORMATION Per g . cat.,
%
Per g. Co,
Catalyst
2 4A 4B 4c 4E
11 660 360 61 49
48 1730 990 200 350
%
Apparent ictivation energy, kcal.
Hydrogen dependence
30 31-33
- 0 . 5 to - 0 . 8 -1.1 to -1.6 -1.2 to -1.6
31-34
- 0 . 7 to -1.2
The table reveals still more forcefully the greater availability of cobalt dispersed in magnesia for this reaction over that in the usual Fischer type catalyst. The apparent activation energy is slightly higher on the magnesia catalysts probably because of the somewhat larger inhibiting effect of the hydrogen. Discussion of Results Comparison with Other Catalysts.-One of the purposes of this study was to place cobalt in its proper position among the metallic catalysts with regard to ability in breaking carbon-carbon bonds in this simple reaction. The table below summarizes results, as nearly comparable as possible, taken from Morikawa’s work on nickel,2 that of Taylor and Jorisg on copper promoted with magnesia, and the present work on cobalt. TABLE 111 Catalyst
Cu-MgO Co-Tho%,etc. Co-MgO-l:5 CO-MgO-5:5 Ni
Initial ratio HI:C2HB
T%mp.,
C.
Rate-% per hour per g. of metal
3:l 2:l 2:l 2:l 1:l
450 255 255 184 184
48 350 15 69
4
It is obvious that cobalt falls between copper and nickel in its ability to catalyze the reaction being studied, and that i t is much nearer to nickel than to copper. This would have been expected from the chemical properties of these elements. Availability of Cobalt for the Fischer Synthesis.-The latter part of the present work, that on the series of cobalt-magnesia catalysts, was undertaken in an attempt partially to elucidate the problem of the availability of cobalt in catalysts for the Fischer synthesis. Because of the relatively high cost of cobalt, i t is of economic importance to find catalysts in which the cobalt is more available for that synthesis, that is, in which the same amount of cobalt shows greater
as an index for the availability of cobalt in catalysts for the Fischer process, and, since i t is a simpler reaction, that i t might aid in putting the whole problem of the Fischer synthesis on a less empirical basis. In this, as in all other catalytic reactions, the only sure method for predicting the properties of a catalyst for the reaction is, of course, to try the reaction on the catalyst. However, since the Fischer synthesis is so complex, it was felt that little fundamental information about catalysts for that process could be obtained by studying that process alone, and that it would be highly desirable to find a simple reaction, involving similar bonds, that would either parallel the Fischer reaction in so far as catalytic activity was concerned, or would provide a method of assessing the availability of the cobalt to carbon-carbon bonds. In the latter case, any effect upon the Fischer catalyst of constituents other than the cobalt should be more obvious, since the hydrogenation-decomposition would indicate the availability to carbon-carbon bonds of the cobalt, leaving to be explained on the basis of promoter action variations in activity for the Fischer process itself. Suppose, for example, that two catalysts were tested in the Fischer process, and that one gave good yields of liquid hydrocarbons and the other poor yields. From that observation alone it would not be clear whether the poor yields in the second case were due to inefficiency of the catalyst in reducing the carbon monoxide to hydrocarbon radicals, or to inefficiency of the catalyst in catalyzing the polymerization of such radicals to the desired higher hydrocarbons. Experiments on the two catalysts with the hydrogenation of ethane would distinguish between the two possibilities, by showing whether there was any difference between the two catalysts in availability of the cobalt to carbon-carbon bonds, and hence to reduction of the carbon monoxide to hydrocarbon radicals. The results summarized in Table I1 show that the availability of cobalt for the hydrogenation of ethane is indeed enhanced by the substitution of magnesia for thoria. On the most active of the cobalt-magnesia catalysts, the rate of the reaction per gram of cobalt was about 36 times as great as that on the cobalt-thoria catalyst, and even on the poorest i t was about four times as great.
509
ETHANOLYSIS OF SPRUCE WOOD
Feb., 1939
All the published information indicates, however, that cobalt-magnesia catalysts do not measure up to the cobalt-thoria catalysts in yield of liquid hydrocarbons. Yields of 150 cc. per cubic meter of gas are now normal in the latter case whereas with cobalt-magnesia catalysts the yields reported are less than 100 cc. of liquid per meter cube. It appears, therefore, that the thoria in a Fischer catalyst exerts a promoter action more specific than that of merely dispersing the cobalt. In other words, good availability of cobalt to carbon-carbon bonds, associated in the Fischer process with the reduction of the carbon monoxide to hydrocarbon radicals, is not a sufficient condition for a good Fischer catalyst. There must be present a promoter] such as thoria, which will give to the catalyst suitable properties, presumably for catalyzing the polymerization of the hydrocarbon radicals to the desired products. More evidence for the specific action of the promoter is found in the type of hydrocarbon product resulting from cobalt catalysts with different promoters. The data in the literature indicate that alumina as a promoter gives rise to a light spirit, thoria to the normal Fischer spirit, and manganese oxide to paraffin wax. In these cases, obviously, the promoter is exercising an influence upon the extent of polymerization that takes place. This suggests that, although it is a better dispersing agent for the cobalt (which undoubtedly initiates the reduction of carbon monoxide t o hydrocarbon radicals), the magnesia is less ef-
[CONTRIBUTION FROM THE
DIVISION OF INDUSTRIAL
ficient than thoria in bringing such radicals into the polymerizing reactions which ultimately yield the Fischer spirit.
Summary 1. The rate of the hydrogenation-decomposition of ethane has been studied on a cobaltthoria-copper-kieselguhr catalyst, and upon four cobalt-magnesia catalysts containing varying proportions of the components. 2. The results indicate that, for this reaction, cobalt is somewhat less active than nickel, but much more active than copper. 3. This reaction on cobalt is inhibited by hydrogen, but to a less extent than the same reaction on nickel. The dependence of the rate on hydrogen concentration varies with the nature of the second component of the catalyst. 4. The cobalt-magnesia catalysts are markedly more active in this reaction than the cobaltthoria-copper catalyst. I n the range of composition studied, the activity of the catalysts of the cobalt-magnesia series increased with increasing cobalt content. 5 . These results have been discussed in connection with the work of previous investigators on nickel and copper. 6. The use of the hydrogenation-decomposition of ethane as an index to the availability of cobalt in catalysts for the Fischer synthesis of higher hydrocarbons from water gas has been discussed. PRINCETON, N. J.
AND CELLULOSE CHEMISTRY,
RECEIVED OCTOBER3, 1938
MCGILLUNIVERSITY]
Studies on Lignin and Related Compounds. XXXV. The Ethanolysis of Spruce Wood BY A. B. C R A M E RM. , ~ J. HUNTER^ AND HAROLD HIBBERT The isolation from extracted lignins of various benzene, l-n-propyl-3-methoxy-4-hydroxybenaromatic products3”such as veratric, dehydrodiver- ~ e n has e ~led ~ to the formation of certain theories atric and isohemipinic acids3b; protocatechuic of lignin structure, which would seem to be based acid and gallic acidaC; catechol, phenol, o-cresol, on unsatisfactory and very incomplete experiguaiacol, creosol, l-vinyl-3-methoxy-4-hydroxy-mental evidence. Up to quite recently there (1) Post-graduate student and holder of Hibbert-Cole Fellowship. has been a general tendency to ascribe a highly (2) Post-graduate student and holder of Dow Chemical Company complex structure] or series of closely related Fellowship. (3) (a) For review see “Chemistry of Lignin,” M. Phillips, Chem. structures, to lignin, but there is now a growing R e v . , 14,103 (1934); (b) Freudenberg a n d co-workers, Ber., 71, 1821 opinion4that lignin may possibly exist in the wood (1938); 69, 1415 (1936); (c) Hibbert and co-workers, THIS JOURNAL, 68, 345, 348 (1936); 69, 2447 (1937); 60, 565 (1938); C a n . J. Research, 1 6 8 , 54, 68 (1938); (dJ Phillips and Goss, I n d . Eng.
Chem., 24, 1436 (1932).
(4) Hilpert and co-workers, Ber., 67, 1157 (1934); 68, 16,371,380, 1576 (1935); 69, 680 (1936); 70, 108, 413, 514 (1937); 2. angew. Chem., 49, 54 (1936); Cellulosechemie, 16, 92 (1935); 17, 124 (1936).
